Positively Transitioned Sentiment Dialogue Corpus

本数据集名为“Positively Transitioned Sentiment Dialogue Corpus”，由华为技术有限公司人工智能应用研究中心创建，专注于收集和分析多轮对话中情感状态从负面或中性向正面转变的数据。数据集包含67,205条对话，每条对话均经过精心筛选，确保情感状态的正向转变。创建过程中，研究团队采用了先进的情感预测模型和表情符号情感映射技术，确保数据的质量和准确性。该数据集主要用于开发情感感知型聊天机器人，旨在提升机器人在处理用户情感问题时的表现，使之更加接近人类水平。

This dataset, named Positively Transitioned Sentiment Dialogue Corpus, was developed by the AI Application Research Center of Huawei Technologies Co., Ltd. It focuses on collecting and analyzing multi-turn dialogue data where the emotional state shifts from negative or neutral to positive. The corpus contains 67,205 dialogues, each carefully screened to guarantee a positive emotional transition. During the dataset construction process, the research team adopted advanced sentiment prediction models and emoji sentiment mapping technologies to ensure the quality and accuracy of the data. This corpus is primarily used for developing sentiment-aware chatbots, with the goal of enhancing the robots' performance in addressing user emotional issues and bringing their capabilities closer to human-level performance.